Cross-Layer Task Scheduling for NOMA-Assisted Satellite Edge Computing

Wang, Ruizhi and Xu, Xiaolong and Cui, Guangming and Bilal, Muhammad and Dai, Fei (2026) Cross-Layer Task Scheduling for NOMA-Assisted Satellite Edge Computing. IEEE Internet of Things Journal. ISSN 2327-4662

Text (Cross-Layer_Task_Scheduling_for_NOMA-Assisted_Satellite_Edge_Computing)
Cross-Layer_Task_Scheduling_for_NOMA-Assisted_Satellite_Edge_Computing.pdf - Accepted Version
Available under License Creative Commons Attribution.
Download (2MB)

Abstract

Ubiquitous, low-latency intelligence at the network edge is central to large-scale Internet of Things (IoT) deployments, yet effectively coordinating communication, computing, and backhaul operations across heterogeneous layers remains challenging. This paper presents a unified cross-layer framework for terrestrial–satellite edge computing IoT systems that integrates Non-Orthogonal Multiple Access (NOMA)-based terrestrial access with local, satellite, and cloud execution. Unlike conventional terrestrial Multi-access Edge Computing (MEC)/edge– cloud scheduling, we jointly optimize partial offloading and path selection over a NOMA-coupled uplink and a multi-hop satellite edge/cloud execution chain under end-to-end latency coupling. The framework jointly determines path selection and partial offloading to minimize a latency–energy objective using accurate end-to-end models. Within this framework, two complementary scheduling algorithms are developed. The Centralized Optimal Cross-Layer Scheduler (COCS) formulates the joint scheduling problem as a mixed-integer nonlinear program (MINLP) with logarithmic and bilinear terms. It employs the spatial branch-and-bound (sBB) method within a commercial solver to obtain a global solution, serving as a performance benchmark. The Decentralized Game-Theoretic Scheduler (DGTS) models user decisions as an ordinal potential game (OPG) and achieves distributed convergence via best-response dynamics (BRD), enabling scalability and adaptability to large networks. Extensive simulations demonstrate that COCS achieves the global optimum while DGTS attains near-optimal performance with much lower complexity. These results validate the effectiveness of the proposed cross-layer framework and highlight the importance of coordinated communication–computation–backhaul design for terrestrial–satellite integrated edge computing.